(2021) 40:475–491 https://doi.org/10.1038/s41388-020-01560-0

REVIEW ARTICLE

Insights from transgenic mouse models of PyMT-induced breast cancer: recapitulating human breast cancer progression in vivo

1,2 2,3 2 1,2,3 Sherif Attalla ● Tarek Taifour ● Tung Bui ● William Muller

Received: 14 August 2020 / Revised: 27 October 2020 / Accepted: 6 November 2020 / Published online: 24 November 2020 © The Author(s) 2020. This article is published with open access

Abstract Breast cancer is associated with the second highest cancer-associated deaths worldwide. Therefore, understanding the key events that determine breast cancer progression, modulation of the tumor-microenvironment and metastasis, which is the main cause of cancer-associated death, are of great importance. The mammary specific polyomavirus middle T antigen overexpression mouse model (MMTV-PyMT), first published in 1992, is the most commonly used genetically engineered mouse model (GEMM) for cancer research. Mammary lesions arising in MMTV-PyMT mice follow similar molecular and histological progression as human breast tumors, making it an invaluable tool for cancer researchers and instrumental in understanding tumor biology. In this review, we will highlight key studies that demonstrate the utility of PyMT derived

1234567890();,: 1234567890();,: GEMMs in understanding the molecular basis of breast cancer progression, metastasis and highlight its use as a pre-clinical tool for therapeutic discovery.

Introduction commercial sources [2, 3]. This model has been used to study several aspects of breast cancer including; initiation, Breast cancer is the most common cancer in women and is histological and molecular progression, metastasis and, responsible for the second highest number of cancer- more recently with the rise of immunotherapies, cancer associated deaths [1]. The development of various Geneti- immunology. Despite not being a human oncogene, PyMT cally Engineered Mouse Models (GEMMs) has been mimics the signaling of receptor kinases which are instrumental in elucidating the molecular events involved in commonly activated in many human malignancies includ- breast cancer initiation, progression and metastasis (Fig. 1). ing breast cancer. The MMTV-PyMT (polyomavirus middle T antigen) In this review, we will discuss the use of PyMT models GEMM (also known as MMTV-PyVmT or MMTV-MT), to assess various angles of tumor progression, highlighting has revolutionized the field of oncomice and, to this day, key studies and findings relating to the human disease. remains the most commonly used GEMM in the field of Finally, we will review several key therapeutic interventions cancer research due to the rapid development of multifocal and how these models were instrumental in their advance- tumors, extensive lung metastasis and its availability from ment to the clinic.

The PyMT GEMMs faithfully recapitulate These authors contributed equally: Sherif Attalla, Tarek Taifour human breast cancer progression

* William Muller First described by our group in 1992, the MMTV-PyMT [email protected] FVB/NJ strain made use of the murine mammary tumor 1 Department of Biochemistry, McGill University, Montreal, QC virus long terminal repeat promoter (MMTV-LTR) to drive H3A 1A3, Canada expression of Polyomavirus Middle T antigen (PyMT) 2 Goodman Cancer Research Centre, McGill University, [2, 4, 5]. Expressing the middle T oncogene in the mam- Montreal, QC H3A 1A3, Canada mary epithelia resulted in rapid transformation and gen- 3 Faculty of Medicine, McGill University, Montreal, QC H3A 1A3, eration of multifocal tumors that readily metastasize to the Canada lungs [2]. Tumors arising from the luminal cells of the 476 S. Attalla et al.

Fig. 1 Notable advances in GEMMs of breast cancer. The unen- field of oncomice in 1997 allowed for conditional ablation within the veloped double stranded was discovered in mammary epithelium. Mutant PyMT strains in 1998 were instrumental 1953. The viral antigen with transformative capacity, the middle T in dissection of PyMT associated signaling. The MTB strain was (PyMT), was identified in 1981. Our lab published the first highly introduced in 2002 which is used to drive expression of the PyMT metastatic mammary tumor model expressing PyMT in the mammary coupled to Cre recombinase in the tetO-MIC published in 2014. gland in an FVB/NJ background. Introduction of MMTV-Cre to the mammary gland progress through distinct histological genes such as ErbB2, Septin9, Col1a1, and Chad are ele- stages that mimic human ductal breast cancer progression vated in PyMT tumors, as is the case in human breast [3]. Nulliparous mice from founder #634 develop hyper- cancers [3, 13, 19]. Although less common, deletions have plasia at 4 weeks of age, which progress through adenomas, been detected in chromosome 4, which codes for several mammary intraepithelial neoplasia (MIN), early and late tumor suppressors [13, 15]. Single nucleotide variants and carcinomas before metastasizing to the lungs, forming activating mutations were detected through whole-exome adenocarcinomas in the lung parenchyma (Fig. 2)[2, 3]. studies and were found to be very common in naïve PyMT profiling revealed that PyMT tumors tumors and drive tumor progression and metastasis [14]. cluster with the luminal B subtype of human breast cancers Moreover, whole genome sequencing detected several [6]. They also display loss of estrogen receptor (ER) and mutations in PyMT tumors [19]. For example, identical progesterone receptor expression (PR), overexpression of mutations in the gene coding for the receptor tyrosine ErbB2 and cyclin D1 as the disease progresses (Fig. 2), phosphatase Ptprh are seen in 81% of PyMT tumors [19]. mimicking human breast cancers with poor prognosis [3, 7]. This mutation renders the phosphatase incapable of Loss of ER is common in breast tumors resistant to endo- dephosphorylating EGFR, causing its constitutive activation crine therapy, as well as in recurrent cancers [8]. End-stage [19, 20]. Similar Ptprh mutations were detected in non- tumors also display robust infiltration by various immune small cell lung cancers and showed sensitivity to tyrosine cells such as macrophages and T cells which contribute to kinase inhibitors [20]. Furthermore, our group has recently progression and metastasis [3, 9–11]. Together these identified functionally active point mutations within the pathological and molecular analyses make this GEMM a Mtor gene in RHEB GTPase deficient PyMT tumors [21]. clinically relevant tool that models the human disease. Significantly, many of these Mtor mutations have also been Several subsequent studies revealed that the expression found in human tumors [21]. It is important to note that the of the PyMT oncogene is coupled to additional biochemical amplifications, deletions and genetic changes start during and genetic events that drive tumorigenesis and metastasis the hyperplasia stage and continue to change as the tumors [12]. To that end, genomic analysis and whole exome advance to later stages, indicating progressive molecular sequencing was performed on the arising tumors [13, 14]. changes that parallel the histological changes in this model While these tumors maintain their diploid status, amplifi- (Fig. 2)[22]. Overall, these findings support the concept cations and deletions have been reported [13]. For example, that full malignant transformation requires additional amplifications of chromosome 11 are common in these genetic events in addition to PyMT’s expression which tumors [13–17]. Chromosome 11 is orthologous to chro- leads to generation of transcriptionally heterogeneous mosome 17 in humans, which is commonly amplified and tumors [19]. The fact these aberrations are also detected in associated with increased PI3K activity due to amplification human breast cancer reinforces the view that tumor pro- of the ErbB2 locus [15–18]. Expression of chromosome 11 gression in this GEMM faithfully recapitulates the many Insights from transgenic mouse models of PyMT-induced breast cancer: recapitulating human breast cancer. . . 477

Fig. 2 Progression of PyMT GEMMs. Both PyMT driven mouse the lung parenchyma. Time points at which 50% of the animals have models progress through the four main stages of cancer, akin to human these pathologies for either MMTV-PyMT or MTB/tetO-MIC in tumors. Tumorigenesis starts with abnormal proliferation and hyper- weeks. Biomarkers associated with breast cancer shown at corre- plasia, which then progresses through adenomas, mammary intrae- sponding time points based off the MMTV-PyMT. Created with pithelial neoplasia (MIN), early carcinomas and culminates in late Biorender.com. carcinomas that metastasizes to the lungs, forming adenocarcinomas in complex stages and heterogeneity of human breast cancer SRC kinases, ShcA, phosphatidylinositol 3-kinase (PI3K) within a very short latency. and phospholipase C-gamma 1 (PLCγ 1) among others [23– 26]. These interactions depend on PyMT’s ability to form a complex with the scaffolding PP2A, which then The PyMT is a potent oncogene that recruits SRC family kinases [27]. PyMT can be phos- associates with several cancer specific phorylated by SRC kinase on tyrosine 315, 322, and 250 signaling pathways [28]. Each event activates a different pathway, which contributes to transformation. Phosphor- The original rationale for studying the transforming poten- ylation at residue 250 recruits the SchA adapter protein that tial of PyMT in the mammary epithelium derived from its can recruit and activate the Ras/MAPK pathway [24, 29]. capacity to activate a number of key cell-intrinsic signaling Tyrosine 315 is known to activate the PI3K pathway while pathways involved in human breast cancer progression. phosphorylation of the 322 residue recruits PLCγ 1[25, 30]. While the PyMT oncogene lacks a kinase domain, it mimics SRC recruitment is essential for PyMT mediated signaling an activated receptor through its association and transformation as full-body knockout of c-Src rendered and interaction with a number of signal transducers such as PyMT incapable of generating mammary tumors [31]. 478 S. Attalla et al.

Subsequent analyses of the full-body knockout c-Src strain Conditional knockouts using MMTV-Cre is a common revealed major dysfunction in ERα signaling, indicating way to bypass embryonic lethality and multiple tissue interplay between epithelial specific and stromal interac- effects. However, selective pressure in MMTV-PyMT/ tions [32]. More recently, in vivo and in vitro epithelial MMTV-Cre bigenics, driven by stochastic expression of the ablation of c-Src revealed that SRC is associated with cell MMTV promoter, allows for generation of escapee tumors cycle progression [33]. [38]. These tumors would express the PyMT oncogene and The importance of the PI3K and SchA signaling path- lack Cre recombinase expression and would therefore retain ways was highlighted by the generation of mutant MMTV- expression of the conditional allele. The generation of such PyMT (Y250F, Y315/322F) strains [34]. Mice expressing escapee tumors renders interpretation of the role of critical Y250F PyMT are incapable of binding SchA while Y315 signaling molecule in tumor induction difficult to assess. To and Y322 mutant impairs PI3K signaling [34]. GEMMs address this issue, a doxycycline inducible model of PyMT expressing the mutant PyMT exhibited a delay in tumor linked to Cre recombinase, referred to as the tetO-MIC onset and impaired metastatic progression [32]. Remark- (tetO-PyMT-IRES-Cre), was generated [39]. Using an ably, a proportion of the metastatic mammary tumors that MMTV driven reverse Tetracycline Transactivator strain arose in PyMT Y250F GEMM possessed either point (MTB) [40], doxycycline-dependent expression of PyMT mutations or in frame deletions that restored the NPTY and Cre recombinase in the mammary epithelial cell can be motif required for ShcA binding, indicating a strong achieved in MTB/tetO-MIC bigenics (Fig. 3). Upon selective pressure for retention of this signaling pathway induction, these mice progress through the same histo- [34]. Consistent with this, crossing MMTV-PyMT and pathological stages seen in MMTV-PyMT and are highly various knock-in models of ShcA mutants resulted in pro- metastatic to the lung [39]. Thus, this model offers the same found impact on tumor induction [35]. Detailed molecular advantages and human relevance of MMTV-PyMT mice analyses revealed that different ShcA mutants correlated with the ability to coincidentally ablate additional genes with differential activation of the downstream STAT1 and with no escapee phenomenon [21, 41, 42]. Another inter- STAT3 transcription factors [36]. In contrast, the PyMT esting facet of the inducible MTB/tetO-MIC strain is that Y315/322F GEMM did not revert back to wild-type form, PyMT is expressed only in the adult animal and therefore, is but rather upregulated the expression of ErbB2 and ErbB3 not recognized as self-antigen but a tumor specific antigen which in turn increased PI3K signaling indicating that that is recognized and targeted by the immune system. In restoration of this key signaling pathways can also occur in this regard, it is interesting to note that, in contrast to the an indirect fashion [34]. original MMTV-PyMT strain where mammary tumor penetrance is 100% [2], although all female MTB/tetO-MIC develop mammary epithelial hyperplasias only 87% of these Second generation inducible strain offers animals develop mammary tumors due to immune-selective temporal and spatial regulation of PyMT pressure [39, 41, 42]. expression, ablation of conditional alleles Further confirmation of the MTB/tetO-MIC model and an in vivo model of tumor induction undergoing active immune surveillance stems from crosses with the conditional Stat3 strain. While epithelial ablation The MMTV-PyMT model has been utilized together with of STAT3 did not impact the initial hyperplastic expansion either full-body knockouts or with mammary specific of the epithelial tree, the nascent lesions were eliminated by expression of Cre recombinase (MMTV-Cre) [37] along massive immune response involving both anti-tumor mac- with conditional alleles to knockout various proto- rophage and T cell populations [41]. Moreover, only 20% and examine their effects on mammary tumor- of mammary gland deficient STAT3 animals developed igenesis and metastasis. While full-body knockout studies focal mammary tumors at a significantly delayed onset were very informative and allowed researchers to elucidate compared to their wildtype counterparts [41]. Interestingly, the roles of different pathways during tumorigenesis, as the STAT3 tumors that escaped immune surveillance failed well as identify novel targets for therapeutic intervention, to metastasize to distal organs, indicating that STAT3 was these studies were very limited and had several drawbacks. critical for the metastatic phase of tumor induction [41]. As highlighted by c-Src studies, germline ablation impacts This key study represents the first GEMM where the various multiple tissues and may even be lethal in some cases phases of immune editing, including tumor elimination, [31, 32]. Many genes involved in cancer progression are immune equilibrium, and immune escape can be mechan- also important for normal mammary gland development, istically dissected [41, 43]. It is important to mention that thus presenting a confounding variable for breast cancer STAT3 is an essential transcription factor that affects a wide studies [32]. variety of cell-intrinsic and extrinsic signaling pathways Insights from transgenic mouse models of PyMT-induced breast cancer: recapitulating human breast cancer. . . 479

Fig. 3 Comparison of PyMT GEMMs. Both strains were initially genetic cross between the MTB (MMTV-rtTA) and the tetO-MIC generated on an FVB/N background. Line #634 is the commonly used strain. Bigenic mice develop multifocal palpable tumors at 22 ± line for the MMTV-PyMT, which develop palpable tumors by 34 ± 7.1 days post doxycycline induction in drinking water. Only 87.1% of 6 days of age at a 100% penetrance. The MTB/tetO-MIC requires a bigneics are expected to develop tumors. Created with Biorender.com. that together contribute to tumor cell proliferation and would be a valuable tool to assess the role of commonly metastasis [44]. PyMT cells lacking STAT3 are capable of seen mutations and their direct contribution to tumorigen- growing in immune-compromised mice [36]. However, esis in a controlled fashion. only around 25% of mice injected with STAT3 deficient PyMT cells develop tumors in an immune-competent background, highlighting STAT3’s essential role in PyMT PyMT GEMMs as a platform to study mediated immune suppression [36]. molecular signaling driving tumor initiation Using primary mammary epithelium from the MTB/tetO- and progression MIC model, 3D organotypic cultures were developed that could be induced to express the PyMT oncogene [45]. Prior The PyMT mouse model has also been extensively used to to doxycycline administration, luminal cells from the glands study the events that promote tumor initiation and mutations were cultured and formed a 3D structure with a lumen, that predispose patients to breast cancer. A comprehensive recapitulating the architecture of the normal mammary review of the effects of different genes on tumor onset and gland [45]. Upon induction of PyMT expression, cells metastasis is provided in Table 1 along with a brief progressively filled the lumen [45]. This progressive filling description of its relevance and impact on the field. Cancer of the duct is akin to tumor induction and progression from initiation describes the early genetic changes that drive a early neoplastic transformation to a ductal carcinoma in situ healthy cell into malignancy. It arises due genetic and epi- in human breast cancers [46]. This provides a tool to dissect genetic (discussed later) changes, mutations, amplifications the early players in mammary gland transformation using an or deletions that inactivate tumor suppressors or activate in vitro model system. This tool can be used as a screening oncogenes. platform for inhibitors, as well as in combination with Deletions in the PyMT mice allowed identification of CRISPR/Cas9 medicated genetic ablation of genes of novel tumor suppressors, whose deletion accelerates tumor interest. Furthermore, MTB/tetO-MIC organoids can be onset and their inhibition would be inadvisable in the clinic. adapted into co-culture systems with fibroblasts or indivi- One such tumor suppressor was the extracellular matrix- dual components of the immune system such as macro- modifying MMP8. Matrix phages in order to study the contribution of these cell types (MMPs) have long been linked to metastasis and tumor to tumor induction and progression. progression due to their role in ECM degradation and Another aspect of the MTB/tetO-MIC strain is that it can cancer cell invasion [48]. MMPs were previously regarded be crossed into conditional, mutant knock-in strains. These as the key to cancer cells’ ability to spread to distant organs knock-in mice could carry a mutant within intronic and several inhibitors were developed and tested in clinical sequences and a loxP site flanking the wildtype exon [47]. trials [49]. Ablation of MMP8 in the MMTV-PyMT model Cre expression allows deletion of the wildtype exon and in- accelerated tumor onset and increased incidence of lung frame expression of the mutant version [47]. These models metastasis [50]. This study was the first to show an in vivo 480 Table 1 Genes affecting onset and progression in PyMT. Gene General function Strain Effect on tumorigenesis Relevance Reference background

HIF1α Transcription factor, important for the ability of FVB/NJ Delayed tumor onset and decreased lung Presented evidence that Hif1α is important for cancer [54, 55] mammalian cells to respond to decreased oxygen metastasis. stem cells, and provided rationale for targeting it to levels. prevent cancer recurrence. CCR6 Chemokine receptor for the ligand CCL20. Found on C57BL/6J Delayed tumor onset, due to defective Macrophages (especially M2) promote tumorigenesis [61] dendritic cells, T cells, B cells and NK cells. initiation and tumor initiation. MMP8 Collagen degrading enzyme. Functions to modify FVB/NJ Accelerated tumor onset Evidence that MMPs may have an anti-cancer role. [50] the ECM MMP8 functions as a tumor suppressor. This is why their inhibitors may not be good in clinic RAG1 Recombination activating gene 1. Essential for VDJ C57BL/6J Accelerated tumor onset Evidence that the adaptive immune system suppresses [43] recombination. Knockout impairs B and T cell tumor formation and plays a role in immune development. Mice lack an adaptive immune system surveillance. IL15 Cytokine important for NK cell differentiation, C57BL/6J Accelerated onset and tumor growth IL15 plays an important role in promoting an anti- [60] proliferation and activation of T cells. tumor immune response. Early on in tumor formation, it stimulates 2 types of innate T lymphocytes, as well as CD8 and NK cells. Evidence of immune surveillance STAT3 Major transcription factor. Regulates expression of FVB MTB/tetO- Delayed tumor onset but no effect on STAT3 is dispensable for tumor initiation but is [41] genes important for survival, proliferation, MIC initiation. Tumor bearing mice were essential for progression and metastasis. Also differentiation and immune supression completely void of metastasis. highlighted role of STAT3 in immune-suppression. The model recapitulates immunoediting TBRII Type II TGFβ receptor FVB/NJ Accelerated tumor onset Absence of TGFβ signaling accelerates tumor [105] formation and showed that TGFβ is a tumor suppressor CSF-1 Cytokine that promotes macrophage differentiation FVB/C57Bl/ No change in tumor onset. Depletion of No change in initiation. Highlighted the role of [11] and recruitment. 6J mixed Tams led to decrease in pulmonary macrophages in promoting breast cancer metastasis metastasis. and progression. MALAT1 Long non-coding RNA that is upregulated in tumors C57Bl/6J No difference in tumor onset but slowed First in vivo evidence to show that this long non- [106] with high chance to metastasize down tumor growth and decreased coding RNA Is not essential for tumor initiation but metastasis. regulates progression and metastasis. ERBB3 Heterodimerizes with ErbB2 to induce proliferation, FVB/NJ Delayed tumor onset and decreased Activation of ErbB3 is required for transformation and [66] growth and survival metastasis thus it is a viable therapeutic target. CTSB Cathepsin B is a lysosomal cysteine protease that FVB/NJ Delayed tumor onset and decreased lung CTSB is important for overcoming tumor dormancy [107] degrades ECM metastasis and promoting tumor growth by degrading the ECM EZH2 Methyltransferase, part of PRC2. Important for gene FVB MTB- Delayed tumor onset and decreased Highlighted context and subtype dependent role of [42] silencing. tetO-MIC metastasis Ezh2 in breast cancer progression GCLM Modifier subunit of glutamate cysteine ligase (GCL), C57BL/6J Delayed tumor onset Shows that synthesis of the antioxidant GSH [108] which synthesizes glutathione. facilitates tumor initiation but is not essential for later al. et Attalla S. stages of cancer progression PTHRP FVB/NJ [109] Insights from transgenic mouse models of PyMT-induced breast cancer: recapitulating human breast cancer. . . 481

evidence of an MMP acting as a tumor-suppressor and that its therapeutic inhibition would not be advisable [50]. ] ] 59 21 Human studies have also shown that higher levels of MMP8 [ [ are correlated with good prognosis and better outcomes in breast cancers [50]. As tumors progress, they become deprived of oxygen, especially at the center of the tumor (Fig. 2)[51]. This phenomenon is known as hypoxia and induces the expres- sion of hypoxia inducible factors (Hifs) [52]. HIF1α,in particular, is overexpressed in several types of cancers and is known to be associated with poor prognosis and decreased survival in cancer patients [53]. The levels of HIF1α progressively increase with tumor progression in the MMTV-PyMT model, mimicking human patients and making it an ideal model to study the roles of hypoxia on breast cancer progression, metastasis and therapy [54]. Mammary epithelial cell deletion of HIF1α in MMTV- PyMT mice significantly delayed tumor onset, slowed down First study to elucidatePTHRP the in tumor-promoting breast role cancers,potential of and prognostic highlighted indicator it or as therapeutic a target. Highlighted Snail1 as aproliferation major and regulator a of potentialinhibit TIC target cancer for recurrence. therapies to Highlighted that mTORC1 signalingPyMT is tumorigenesis. essential Arising for mutationsuse highlight of the PyMT modelsmodulation in as recapitulating human genetic tumors. tumor growth and caused a significant reduction in lung metastases [54, 55]. This study highlighted the role of this oxygen-dependent transcription factor in tumor initiation and progression to advanced disease. Deletion of HIF1α in vitro decreased the tumorigenic potential of PyMT cells, as well as their ability to form tumor-initiating cells (TICs) [54]. TICs and cancer stem cells are major concerns for breast cancer patients and are a major cause for cancer recurrence [56]. Cancer stem cells have been characterized in the PyMT models of breast cancers, commonly expres- sing markers such as CD24+CD29+CD49f+Sca-1lo [57]. Effect on tumorigenesisDelayed tumor onset, progressiondecreased and metastasis Relevance Reference Delayed tumor onset dueproliferative to capacity decreased of TIC Delayed onset due toof reduced mTORC1 activity Moreover, the transcription factor GATA3 was identified as a tumor suppressor that inhibits proliferation of cancer stem cells [58]. Knockout of Gata3 accelerates tumor onset due to the increased TIC capacity of luminal progenitor cells [58]. On the other hand, deletion of the zinc-finger tran- background FVB/NJ/C57Bl/ 6J mixed FVB/NJ MTB/ tetO-MIC scription factor Snail1 delayed tumor onset and caused a significant reduction in the proliferative capacity of TICs [59]. Thus, these studies highlight a dynamic balance between transcription factors that regulate TIC and show that PyMT models are ideal for the study of cancer stem cells and develop interventions against them that could prevent relapse and cancer recurrence. In addition to TICs and cancer stem cells, the MTB/tetO- MIC strain, in particular, is a valuable tool to study tumor recurrence. Tumor-bearing mice stripped of doxycycline drinking water and provided with normal drinking water, show strong robust tumor regression to a virgin-like state [39]. However, these animals do eventually display doxycycline-

nger Transcription factor, known to independent tumor recurrence [39]. Recurrent tumors displayed fi various pathologies not seen in the pre-regression tumors and Parathyroid hormone related protein, afound secreted factor to be elevatedrange in of breast developmental cancers. functions Itgrowth. and has stimulates a wide regulate EMT mTORC1 signaling infrequent mutations in TP53 [39]. Thus, the MTB/tetO-MIC

(continued) strain may be a valuable tool to study recurrent tumors that do not respond to the primary therapy since most recurrent tumors SNAIL1 Zinc Table 1 Gene General function Strain RHEB1 GTPase essential for activation of are no longer dependent on PyMT expression. 482 S. Attalla et al.

PyMT mouse models have also been used to study the These studies demonstrated that loss of β1 integrin in an roles of the immune system in cancer initiation and pro- emerging PyMT tumor resulted in a dramatic impairment of gression (see later). A germline knockout of RAG1, a key mammary tumorigenesis [38]. Significantly, the PyMT enzyme in VDJ recombination, eliminates the adaptive tumors that emerged retained expression of β1 integrin due immune system (B and T cells) and accelerates tumor onset to escapee from Cre mediated recombination (see earlier) in the MMTV-PyMT (C57BL/6J) mouse [43]. Knocking highlighting the importance of β1 integrin in tumorigenesis. out the interleukin IL15 also caused a similar acceleration in In addition to impacting the initiation phase of tumor pro- tumor onset [60]. Analysis of early lesions revealed that gression, ablation of β1 integrin function in PyMT cell lines IL15 stimulates two separate populations of T cells, as well resulted in cell cycle arrest, inducing a state of cellular as natural killer and dendritic cells to target cancerous dormancy [38]. Consistent with the importance of integrin growth, thus highlighting it as a key tumor suppressive signaling in mammary tumor progression, mammary epi- cytokine [60]. On the other hand, macrophages play a dual thelial deletion of the integrin associated signaling protein, role in cancer initiation. While depletion of total macro- Focal Adhesion Kinase (FAK) resulted in a similar impact phages through the knockout of Colony Stimulating Factor on PyMT tumor progression and induction of cellular dor- 1 (CSF1) had no effect on tumor onset, the depletion of mancy [68]. Again, like β1integrin ablation studies, PyMT alternatively activated macrophages (M2 polarized) through tumors that arose retained FAK expression due to escapee knockout of CCR6 delayed the onset of tumors [11, 61]. As from Cre mediated recombination [68]. Future studies using discussed in the previous section, deleting STAT3 in the the MTB/tetO-MIC model with these key regulators of MTB/tetO-MIC model results in immune-system remodel- tumor dormancy will provide critical insight into the ing and complete elimination of hyperplastic growth, dra- molecular and cellular mechanisms governing emergence matically delaying tumor onset [41]. Collectively, these from tumor dormancy. studies highlight the diverse and dynamic roles that cells of the innate and the adaptive immune system play in cancer initiation and progression. Further analysis is required to PyMT tumor progression is associated with fully understand the complex interactions between cancers epigenetic modulations and different populations of immune cells, which will allow for development of more effective immunotherapies. In addition to mutations, amplifications and other genetic About 70% of breast cancers are ER positive, even at the changes, epigenetic alterations have also been shown to be metastatic setting [62]. The MMTV-PyMT has been used to common in cancers, maybe even driving tumor development assess the functionality of the ER. PyMT tumors are sen- through epigenetic silencing of tumor suppressor or inducing sitive to the estrogen receptor modulator, tamoxifen, despite oncogene expression [69]. Epigenetic modifications are the low expression of nuclear ER in end stage tumors [63]. reversible changes that affect gene expression but cause no PyMT animals lacking the ER co-transcription factor change to the DNA sequence. They include histone mod- Activated In Breast cancer 1 (AIB1) display delayed tumor ifications and DNA methylation among others. DNA onset and a reduction in pulmonary metastasis [64]. This methylation is linked to gene silencing and hyper-methylation suggests that the metastatic cascade of PyMT is driven, at of various tumor suppressor promoter regions is common in least in part, by the ER [64]. Moreover, exogenous sup- cancers [70]. Histone methylation by the Polycomb Repres- plementation of estradiol to mice harboring PyMT tumors sive Complex 2 (PRC2) is a major mechanism of gene increases rate of tumor growth which was associated with silencing [71]. EZH2 is the methyltransferase in the PRC2 increased vasculature organization [65]. On the other hand, complex, responsible for adding three methyl groups onto similar to ~30% of breast cancers, advanced PyMT tumors lysine 27 of histone 3 (H3K27me3) [71]. Elevated expression express high levels of ErbB2 [3] Indeed, PyMT tumors of EZH2 is seen in breast cancers and has been linked to poor respond to the EGFR/ErbB2 tyrosine kinase inhibitor prognosis [72]. Overexpressing EZH2 under the control of the lapatinib [66]. Furthermore, knockout of the ErbB3, the MMTV promoter (MMTV-EZH2) resulted in the formation preferred heterodimerization partner of ErbB2, delayed of mammary gland hyperplasia and adenomas, the first stage tumor onset and impaired metastasis of MMTV-PyMT [66]. in breast cancer progression [73]. This study highlighted that Ultimately, the cancer cell senses its extracellular matrix EZH2 may be important for tumor initiation and the pro- environment through multiple receptors, including integrin gression of normal epithelium to hyperplasia [73]. Gene receptors, which in turn activates oncogenic downstream expression and DNA methylation status were assessed in signaling pathways within the cancer cell [67]. The critical glands of MMTV-PyMT animals at various stages of disease importance of integrin receptors in mammary tumor pro- progression [74, 75]. Analysis revealed an increase in dif- gression was highlighted through mammary epithelial ferentially expressed genes, as well as differentially methy- ablation of the β1 integrin (CD29) in MMTV-PyMT [38]. lated cytosines as the disease progresses [74, 75]. Insights from transgenic mouse models of PyMT-induced breast cancer: recapitulating human breast cancer. . . 483

Interestingly, hypermethylation of promoters, which was gene, which encodes a GTPase activated protein, harbored a associated with reduced gene expression, were enriched for in threonine to alanine substitution in the lower metastatic the late stage carcinomas compared to earlier hyperplasia [74]. strain, which increased SIPA1 binding to its negative reg- This was also coupled to an increased expression of PRC ulator, AQP2 [81]. Further support of the relevance of this components including EZH2 [74]. These studies highlight GEMM was that SIPA1 expression is associated with that several epigenetic players, DNA and histone methyla- metastatic progression of human prostate cancer [81]. It is tions, are associated with tumor progression and may act in worth noting that MMTV-PyMT FVB/NJ × AKR/J is a synergy to suppress expression of tumor suppressors. Func- highly metastatic strain, whereas MMTV-PyMT FVB/NJ × tionally, conditional ablation of EZH2 in the MTB/tetO-MIC DBA/2J is a low metastatic strain which are associated with model significantly delayed tumor onset and decreased the the polymorphisms of Sipa1 [78, 79]. incidence of lung metastasis, showing that EZH2 is essential for progression of mammary tumors and metastasis [42]. This phenotype was also recapitulated through pharmacological The role of the host tumor inhibition of EZH2 [42]. Off note, treatment of an ErbB2 microenvironment in mammary driven mouse model with an inhibitor of EZH2 revealed tumorigenesis: Implications for disease strong synergy with anti-ErbB2 targeted therapy which progression and metastasis highlights the importance of EZH2 in several cancers and emphasizes its clinical relevance [76]. Mammary gland density

In addition to providing genetic and mechanistic insight into The genetic background of MMTV-PyMT tumor initiation, PyMT driven mouse models have also influences tumor onset and metastasis been used to identify and study various risk factors for developing breast cancers. For several decades, breast tissue Mouse genetic background plays a fundamental role in the density has been linked to breast cancer incidence in women biology of the tumor. FVB/NJ MMTV-PyMT tumors arise [82, 83]. While the evidence was mostly correlational, it significantly faster than in C57Bl/6J MMTV-PyMT [77]. was observed that higher breast density is associated with Furthermore, identical knockouts in these different back- higher breast cancer risk and that hyperplasia and ductal grounds give rise to different results. Prominently, iNOS carcinomas in situ often begin in the densest parts of the knockout only shows an effect on tumor onset and metas- breast [84]. The first causal link of breast density to cancer tasis in the C57Bl/6J but not the FVB/NJ background [77]. incidence came using the MMTV-PyMT mouse carrying a A similar effect was shown when RAG1 was knocked out mutation that makes collagen resistant to degradation by in both FVB/NJ and C57Bl/6J, with only the C57Bl/6J collagenases (Col1a1tm1jae)[85]. This mutation resulted in strain showing accelerated onset [10, 43]. Interestingly, increased collagen deposition in the breast, mimicking high- MMTV-PyMT FVB/NJ was crossed one generation into 27 density human breast tissue [85]. The increased tissue inbred mouse background strains, 13 of which displayed a density was the stimulating event driving tumor formation, significant reduction in pulmonary metastasis and 10 had increasing the number of tumors per mouse, as well as altered tumor kinetics compared to the parental FVB/NJ number of metastatic lesions [85]. This was the first evi- background [78]. This study supported the role of metastatic dence linking breast tissue density to breast cancer inci- modifier genes, potential tumor suppressors or proto-onco- dence and verifying it as a major risk factor for breast genes, that play an important role in tumorigenesis and cancer development. Further analysis found this increased metastasis [78, 79]. Polymorphisms that affect gene tumorigenesis to be mediated by COX2, raising the possi- expressions or silent mutations in genes that display slightly bility of COX2 inhibitors as therapeutic options for breast altered activity levels during neoplasia in these inbred cancer patients with high breast tissue density [86]. strains are sufficient to alter metastatic burden. This could at least in part explain the biological differences between the Macrophages FVB/NJ and C57Bl/6J backgrounds discussed above. Such loci potentially play a role in generating inter-patient tumor Macrophages are one of, if not, the most abundant immune heterogeneity and might allow identification of metastasis cells in the tumor microenvironment [87]. As tumors grow susceptible patients (metastasis-predictive gene expression in MMTV-PyMT mice, a decrease in the mammary tissue signatures), even from premalignant, normal tissue, and macrophages (MTMs) (MHCIIhi, CD11bhi) coupled with an infiltrating non-neoplastic cells [79]. Mtes1 is one metastatic increase in tumor associated macrophages (TAMs) modifier locus identified to be different between the inbred (CD11blo, MHCIIhi, F4/80+, CD64+, MerTK+)is strains [80]. Particularly within the Mtes1 locus the Sipa1 observed [87–89]. Macrophages promote nearly all the 484 S. Attalla et al. stages of cancer metastasis such as preconditioning the pre- epidermal growth factor (EGF) which binds the EGF metastatic niche, supporting tumor invasion, cell extra- receptor found specifically on the cancer cells [9]. In turn, vasation and colonization of the secondary organ [89]. On PyMT cells secrete CSF1, which binds the CSF1 receptor the extreme scale, macrophages can be polarized to either on macrophages promoting their taxis towards the tumor classically activated macrophages (M1), which is associated cells [9]. Imaging revealed that migratory cells leaving the with response to interferon gamma and lipopolysaccharides primary tumor were almost exclusively PyMT carcinoma [87, 89]. These macrophages are associated with production cells and macrophages [9]. Depletion of macrophages, of pro-inflammatory cytokines and antigen presentation using the CSF1 knockout, reduced the number of migratory [87–89]. This population would be associated with cells [9]. Another study knocked out S1P receptors speci- improved prognosis. On the other end of the scale, macro- fically in TAMs using an F4/80-Cre strain, revealed defects phages can be alternatively activated (M2), which are in lymphangiogenesis and corroborated a reduced pul- associated with anti-inflammatory cytokines and promote monary metastatic burden [92]. This data argues that PyMT tissue remodeling, supporting cancer progression and not only metastasize via the hematogenous route (see metastasis [87–89]. Despite popular belief, data arising angiogenesis section) but also via the lymphatics. from characterization of TAMs in MMTV-PyMT revealed that they are not strictly M2-polarized [88]. Mammary tis- T-lymphocytes sue macrophages in untransformed glands, MTMs, resem- bled M2 macrophages more than TAMs [88], which would There is also compelling evidence supporting the role of be consistent with their role of tissue remodeling in a very adaptive immune cells, such as T cells, during PyMT tumor dynamic organ such as the murine mammary gland. progression. The number of CD4+ T cells along with F4/80 Franklin et al. proposed a model in which CCR2+ mono- + macrophages increases during progression [10]. Deple- cytes develop into TAMs, with Vcam1+ expressing cells tion of CD4+ T cells was associated with reduced number being mature TAMs in MMTV-PyMT [87, 88]. None- of circulating tumor cells and pulmonary metastasis [10]. theless, microarray analysis of transcripts expressed in Furthermore, CD4+ deficient lesions expressed elevated phagocytic myeloid cells in end stage MMTV-PyMT levels of M1 macrophage cytokines (TNFα and Nos2) [10]. tumors reveal expression of genes associated with mediat- Further analysis revealed that IL-4 released by CD4+ ing immune responses and angiogenesis [90]. Presence of T cells was capable of polarizing macrophages into more Vcam1+ TAMs was correlated with an increased number of M2-like state while suppressing the M1-like phenotype and exhausted T cells (PD1+Gzmb−CD8)+, highlighting the promoting macrophage expression of EGF, which estab- crosstalk between the innate and adaptive immune respon- lishes the paracrine loop between tumor cells and macro- ses in modulation of the tumor immune response [88]. phages [10, 10, 87, 89]. The role of macrophages in tumor progression and The role of T cells is not limited to the primary site, but is metastasis was highlighted using a null mutant of CSF1, a also implicated at the metastatic niche. PyMT tumors macrophage growth factor [11]. Homozygous null CSF1 overexpressing macrophage stimulating protein (PyMT- MMTV-PyMT mice exhibited a remarkable decrease in MSP) show more local invasive behavior and lead to the tumor progression to advanced disease and lacked evidence of development of osteolytic bone metastasis, not seen in the pulmonary metastasis [11]. This was associated with a dra- parental PyMT mice [93]. This highlights the versatility of matic decrease in F4/80+ macrophages in the primary tumor PyMT models, as additional oncogenes may be over- [11]. This phenotype was rescued using an overexpression expressed or knocked-in (or tumor-suppressor knockouts) CSF1 animal [11]. Furthermore, overexpression of CSF1 in that promote metastatic-organ tropism. Generation of such heterozygous null CSF1 MMTV-PyMT animals accelerated models is important to allow development of therapeutics progression to advanced disease and accelerated the emer- that appropriately target metastatic disease, which is not gence of pulmonary metastasis [11]. Again, this was coin- limited to lungs. ciding with an elevated number of F4/80 cells in the primary MSP binds its receptor, Ron (MST1R), which is tumors [11]. The delayed progression in CSF1 null animals expressed on a variety of cells including breast cancer cells was also rescued by epithelial overexpression of VEGF [91], [94]. A synergistic transplant model of PyMT-MSP cells in suggesting that macrophages may induce VEGF expression in animals harboring mutant Ron receptors lacking tyrosine PyMT cells. kinase ability (Ron TK−/−), demonstrated that the effect is Using an in vivo migration assay and multiphoton ima- driven by the host environment [95]. PyMT-MSP harboring ging of migratory cells revealed a paracrine loop that exists Ron TK−/− animals had elevated splenic CD8+ T cells between PyMT carcinoma cells and macrophages [9]. and elevated CD8+ cells in the tumor margin. CD8+ cells Macrophages enhance cell migration via expression of also infiltrated the metastatic lung and expressed more Insights from transgenic mouse models of PyMT-induced breast cancer: recapitulating human breast cancer. . . 485

TNFα [95]. CD8+ cells from Ron TK−/− tumors had dramatically [99]. Myeloid derived macrophages are key for improved cytotoxic ability [95]. Antibody depletion of CD8 PyMT progression and metastasis (discussed previously) + cells or the use of a SCID mouse rescued the metastatic and are a major source of VEGF [87–89]. However, defect of Ron TK−/− animals [95]. Consistent with this, knockout of VEGF in the myeloid compartment through the pharmacological inhibition of Ron prevented formation of LysM-Cre strain led to faster tumor progression to malig- overt metastasis in the lung post metastatic colonization nancy compared to their wildtype counterparts [99]. This [95]. Further studies have also revealed that suppression of interesting phenomenon might be explained, at least par- the MSP-RON axis potentiates the effect of anti-CTLA4 tially, by the time of anti-VEGF treatment/ablation. Mice checkpoint blockade via crosstalk between macrophages harboring LysM-Cre/VEGF conditional are depleted of and cytotoxic T cells in MMTV-PyMT primary tumors, as macrophage derived VEGF early on during tumor initiation, well as at the metastatic lung [96]. potentially allowing for adaptation mechanism to come into Overall, these studies support the presence of crosstalk play. Another study neutralized VEGF receptor 2 by DC101 between the innate and adaptive immune system that con- and noted a reduction in growth [100]. Analysis of the tributes to metastasis. Furthermore, it raises the possibility immune microenvironment of these tumors revealed a sig- of targeting the immune microenvironment at the metastatic nificant increase in IFN γ producing cytotoxic T cells [100]. site, which would be important in improving therapeutic However, this was associated with an increase in PDL-1 regimens for patients. Studies such as these are possible due expressing PyMT tumor cells as a negative feedback to the high penetrance of metastasis of the PyMT models mechanism [100]. It is therefore reasonable to suggest that within an immunocompetent background. VEGF knockout in macrophages results in elevated expression of tumoral PDL-1 during initiation, which leads Neutrophils to immune evasion and supports the faster growing tumors. Indeed, co-treatment of PyMT tumors with both DC101 and Another subset of immune cells that play a role during anti-PDL-1 was sufficient to restrict tumor growth and was cancer progression and metastasis are the neutrophils. In the associated with vasculature normalization [100]. Further- pre-metastatic lung of MMTV-PyMT mice, an influx of more, neutralization of angiopoietin 2, a secreted endothe- neutrophils (CD11b+Ly6G+) occurs and their numbers lial cell growth factor, by a monoclonal antibody (3.19.3) continue to increase during the metastatic progression [97]. impeded tumor growth in MMTV-PyMT mice at both early Animals with depleted neutrophils display a reduction in and late stages of progression [101]. This neutralization led spontaneous pulmonary metastasis [97]. Moreover, the to vascular regression and increased pericyte coverage of CD11b+LyGhi cell population had elevated expression of the vessels, increased tumor hypoxia, necrosis and extracellular matrix chondroitin proteoglycan, versican, enhanced recruitment of MRC1+(CD206) tumor associated which promotes growth of the metastatic lesions and macrophages [101]. Enhanced recruitment of macrophages mesenchymal to epithelial transition which is required for was also noted post treatment using combretastatin A4 development of the metastasis [97]. phosphate, a vascular disrupting agent, suggesting it as a general mechanism in attempt to rescue the tumor vascu- Natural killers lature after therapy [102]. Interestingly, tumors treated with 3.19.3 did not display rebound angiogenesis despite ele- Natural killer (NK) cells also play an important role in vated levels of VEGFa and MRC1+ macrophage infiltrate, immune surveillance [98]. Genetic and antibody depletion highlighting a key role that angiopoietin 2 may play in of natural killer cells increased pulmonary metastatic burden promoting vascularization in PyMT driven tumors [101]. in PyMT animals [98]. Intranasal IL-12 treatment to pro- Future application of this monoclonal therapeutic approach mote activation of NK cells, reduced pulmonary metastasis, may have important implications on breast cancer treat- emphasizing their role in the metastatic niche and not just in ments. The mechanisms by which PyMT cells model the the primary tumor [98]. microenvironment and suppress the anti-cancer immune response to promote metastasis are summarized in Fig. 4. Angiogenesis Despite the focus on the hematogenous route of metas- tasis, the PyMT model does metastasize, albeit to a low Another key tumor microenvironmental cue is the process extent, to the lymph node, a common metastatic site in of angiogenesis. As lesions of MMTV-PyMT progress to human breast cancer [103]. Lymph node metastasis was early carcinomas, they are associated with increased vas- significantly increased in animals lacking expression of the cular density [99]. As the tumors progress to malignancy, cytochrome P450 Cyp2c44, similar to MMP8 suggesting its the length of the vessels and their tortuosity increases inhibition is inadvisable [103]. 486 S. Attalla et al.

Fig. 4 TME factors that promote metastasis in PyMT. a At the killing ability. PyMT also suppress cytotoxic T cell activity indirectly primary tumor, PyMT cancer cells secrete a range of immunosup- via release of MSP which binds the RON receptor on macrophages pressive and chemotactic molecules. CD4+ T cells recruited to the which leads to suppression of cytotoxic T cell activity. c PyMT cancer primary tumor secrete IL-4 which polarizes macrophages to the cells metastasize via hematogenous route along with macrophages to M2 state. Polarized macrophages in turn release EGF which promotes reach the lung. At the metastatic site, neutrophils release versican cancer cell growth and migration. Myeloid and cancer cell derived which supports growth of the PyMT cancer cells and promotes VEGFA 6 promotes angiogenesis. PyMT cancer cells release CSF1 mesenchymal to epithelial transition. Within the lung microenviron- which promote taxis of macrophages. b Within the primary tumor, ment, PyMT cancer cells supress the cytotoxic T cells and natural PyMT cancer cells suppress activity of cytotoxic T cells both directly killer (NK) cells directly and indirectly. Created with Biorender.com . via PD1/PDL1 and CTLA4/CD80 interactions, inhibiting their tumor

PyMT GEMMS as a preclinical platform for have also been tested in combination with other treatments, therapeutic testing chemotherapies and radiation similar to what physicians prescribe their patients following breast cancer diagnosis. In addition to providing valuable genetic and mechanistic Thus, PyMT models are clinically relevant mouse models insights into breast tumorigenesis, PyMT-driven GEMMS used for pre-clinical testing of various therapeutic methods have been a cornerstone for pre-clinical development and to determine doses, safety, and efficacy. testing of potential therapies. From chemotherapies to tar- geted therapies, immunotherapies and cancer vaccines, all have been tested for their efficiency and dosage using Conclusions PyMT mouse models. Table 2 highlights some of the therapies tested in PyMT models grouped based on the drug Overall, we highlighted several key studies that made use of type. It is important to note that several of these therapeutics either PyMT GEMM that not only provided insight into nihsfo rngncmuemdl fPM-nue ratcne:rcptltn ua ratcne...487 . . cancer. breast human recapitulating cancer: breast PyMT-induced of models mouse transgenic from Insights Table 2 Therapeutics with clinical potential tested in PyMT. Drug name Target/mechanism of action Effect Reference

Immunotherapies ANTI-PDL1 Blocks the Programmed death ligand (PDL1) that is commonly Increased CD8+ T cell activity, decreased tumor burden. [100] expressed by tumor cells to inhibit T cell activity. ADIL12– B7-1 Adenovirus expressing the cytokine IL12 and the co-stimulatory Complete regression of tumors with no detected relapse. At lower dose, [110] molecule B7-1 that activates T cells. it delayed tumor growth and was still effective. ANTI-IL4 Monoclonal antibody against the secreted cytokine IL4. No change in tumor onset, latency or burden. But significantly [90] Antibody results in increase M1 macrophages polarization. decreased the number of lung metastases. Drugs targeting DNA or RNA EZN-3920 ErbB3 antisense Significant reduction in tumor growth and inhibits tumor progression. [66] Targeted therapies GSK-126 Ezh2 methyltransferase inhibitor Decrease in hyperplasias and delay in tumor onset. Complete block of [42] metastasis Buthionine-[S, R]-sulfoximine (BSO) Inhibitor of GSH (Glutathione) synthesis. Early treatment increased ROS levels, reduced tumor burden and [108] disrupted progression to later stages. Treatment upon tumor onset did not change ROS levels and had no effect on tumor burden, suggesting a reason why clinical trials have failed. Anti-MMR nanobody Nanobody targeting CD206+ TAMs Specially binds to and label TAMs, allowing for imaging of tumor [111] stroma and hypoxic regions. Anti-ANG2 antibody (3.19.3) Blocks ANG2’s ability to bind Tie2, inhibiting an important pro- Interfered with angiogenesis, disrupted blood vasculature and inhibits [101] angiogenic axis tumor progression ANTI-CSF1 Inhibits colony stimulating factor, a macrophage differentiation Depleted macrophages and delayed tumor regrowth following radiation [112] factor. PLX3397 Competitive ATP inhibitor for CSF1 receptor Eliminated tumor associated macrophages and delayed tumor growth [112] after radiation DC101 antibody anti-VEGF Reduced tumor growth due to impairment of angiogenesis. It also [100] increased the levels of PD-L1, sensitizing tumors for combination with anti-PDL1 antibody. This combination led to significant reduction in tumor burden. Lapatinib Tyrosine kinase inhibitor (ErbB2 and EGFR) Delay tumor growth, [66] Tamoxifen Selective estrogen receptor modulator (SERM) that inhibits ER Inhibited tumor growth and decreased tumor volume. [63] signaling. Chemotherapy and radiotherapy 5 GY focal gamma irradiation from Ionizing radiation Tumor regressed, but regrowth occurs after a period of stasis. [112] cesium source Doxorubicin Chemotherapy. Inhibits topoisomerase 2, an enzyme important Reduced tumor volume due to necrotic cancer cell death. [113] for DNA replication. 488 S. Attalla et al. how the model functions and behaves, but also provided References insight into clinically relevant cancer biology. Many find- ings from the PyMT models are directly applicable to those 1. Brenner DR, Weir HK, Demers AA, Ellison LF, Louzado C, Shaw of human cancers supporting that these models are clinically A, et al. Projected estimates of cancer in Canada in 2020. CMAJ. 2020;192:E199–205. https://doi.org/10.1503/cmaj.191292. relevant. Despite the rising popularity of human xenografts, 2. Guy CT, Cardiff RD, Muller WJ. Induction of mammary tumors whether of established cell lines or from a patient, these by expression of polyomavirus middle T oncogene: a transgenic models make use of immune compromised mice, which are mouse model for metastatic disease. Mol Cell Biol. – unable to mount an adaptive immune response [104]. This 1992;12:954 61. https://doi.org/10.1128/mcb.12.3.954. 3. Lin EY, Jones JG, Li P, Zhu L, Whitney KD, Muller WJ, et al. makes pre-clinical testing of drugs that make use of the Progression to malignancy in the polyoma middle T oncoprotein adaptive immune system impossible. Moreover, in these mouse breast cancer model provides a reliable model for human models, human cells are growing in a murine micro- diseases. Am J Pathol. 2003;163:2113–26. https://doi.org/10. environment which raises the issue of species incompat- 1016/S0002-9440(10)63568-7. 4. Gross L. A filterable agent, recovered from Ak leukemic extracts, ibilities and a foreign microenvironment [104]. causing salivary gland carcinomas in C3H mice. Proc Soc Exp Co-transplant of human stroma (fibroblasts) has also Biol Med. 1953;83:414–21. become popular but they do not address the issue of 5. Treisman R, Novak U, Favaloro J, Kamen R. Transformation of metastasis as the seeded metastasis is now associated with rat cells by an altered polyoma virus genome expressing only the middle-T protein. Nature. 1981;292:595–600. murine stroma [104]. In addition, xenografts cannot be used 6. Pfefferle AD, Herschkowitz JI, Usary J, Harrell JC, Spike BT, to assess tumor onset as they come from established tumor Adams JR, et al. Transcriptomic classification of genetically lines and their implantation into the fat pad does not engineered mouse models of breast cancer identifies human accurately represent tumor progression as there is no subtype counterparts. Genome Biol. 2013;14:R125. https://doi. org/10.1186/gb-2013-14-11-r125. myoepithelial layer surrounding the luminal cells. Further- 7. Lapidus R, Nass S, Davidson N. The loss of estrogen and pro- more, genetic manipulation of the tumor as a whole is gesterone receptor gene expression in human breast cancer. J currently not feasible. Therefore, the use of PyMT derived Mammary Gland Biol Neoplasia. 1998;3:85–94. GEMMs to model breast cancer remains just as valuable as 8. Lopez-Tarruella S, Schiff R. The dynamics of estrogen receptor status in breast cancer: re-shaping the paradigm. Clin Cancer Res. they ever were. 2007;13:6921–5. https://doi.org/10.1158/1078-0432.CCR-07-1399. 9. Wyckoff JB, Wang Y, Lin EY, Li J, Goswami S, Stanley ER, et al. Acknowledgements SA is supported by the McGill faculty of medi- Direct visualization of macrophage-assisted tumor cell intravasa- cine Gosselin Graduate Studentship. TT is supported by the MD/PhD tion in mammary tumors. Cancer Res. 2007;67:2649–56. https:// stipend awarded by McGill faculty of medicine. WJM is supported by doi.org/10.1158/0008-5472.CAN-06-1823. the Canada Research Chair of Molecular Oncology. This work was 10. Denardo DG, Barreto JB, Andreu P, Vasquez L, TawfikD, also supported by Terry Fox Research Institute Program Group Grant, Kolhatkar N, et al. CD4+ T cells regulate pulmonary metastasis National Institute of Health Research, Canadian Institute of Health of mammary carcinomas by enhancing protumour properties of Research the Canadian Cancer Society Impact Grant and Cancer macrophages. Cancer Cell. 2009;16:91–102. Research Society. 11. Lin EY, Nguyen AV, Russell RG, Pollard JW. Colony- stimulating factor 1 promotes progression of mammary tumors – Compliance with ethical standards to malignancy. J Exp Med. 2001;193:727 40. 12. Maglione J, Moghanaki D, Young LJ, Manner CK, Ellies LG, Joseph SO, et al. Transgenic polyoma middle-T mice model fl fl Con ict of interest The authors declare that they have no con ict of premalignant mammary disease. Cancer Res. 2001;61:8298–305. interest. 13. Ben-David U, Ha G, Khadka P, Jin X, Wong B, Franke L, et al. The landscape of chromosomal aberrations in breast cancer mouse Publisher’s note Springer Nature remains neutral with regard to models reveals driver-specific routes to tumorigenesis. Nat Com- jurisdictional claims in published maps and institutional affiliations. mun. 2016;7:12160 https://doi.org/10.1038/ncomms12160. 14. Ross C, Szczepanek K, Lee M, Yang H, Qiu T, Sanford JD, et al. Open Access This article is licensed under a Creative Commons The genomic landscape of metastasis in treatment-naïve breast Attribution 4.0 International License, which permits use, sharing, cancer models. PLoS Genet. 2020;16:e1008743. https://doi.org/ adaptation, distribution and reproduction in any medium or format, as 10.1371/journal.pgen.1008743. long as you give appropriate credit to the original author(s) and the 15. Montagna C, Lyu M, Hunter K, Lukes L, Lowther W, Reppert T, source, provide a link to the Creative Commons license, and indicate if et al. The Septin 9 (MSF) Gene is amplified and overexpressed in changes were made. The images or other third party material in this mouse mammary adenocarcinomas and human breast cancer cell article are included in the article’s Creative Commons license, unless lines. Cancer Res. 2003;63:2179–87. indicated otherwise in a credit line to the material. If material is not 16. Hodgson JG, Malek T, Bornstein S, Hariono S, Ginzinger DG, included in the article’s Creative Commons license and your intended Muller WJ, et al. Copy number aberrations in mouse breast use is not permitted by statutory regulation or exceeds the permitted tumors reveal loci and genes important in tumorigenic receptor use, you will need to obtain permission directly from the copyright tyrosine kinase signalling. Cancer Res. 2005;65:9695–704. holder. To view a copy of this license, visit http://creativecommons. 17. Rennhack J, To B, Wermuth H, Andrechek ER. Mouse models org/licenses/by/4.0/. of breast cancer share amplification and deletion events with Insights from transgenic mouse models of PyMT-induced breast cancer: recapitulating human breast cancer. . . 489

human breast cancer. J Mammary Gland Biol Neoplasia. 35. Ursini-Siegel J, Hardy WR, Zuo D, Lam SH, Sanguin-Gendreau 2017;22:71–84. https://doi.org/10.1007/s10911-017-9374-y. V, Cardiff RD, et al. ShcA signalling is essential for tumour 18. Orsetti B, Nugoli M, Cervera N, Chuchana P, Ursule L, Redon progression in mouse models of human breast cancer. EMBO J. R, et al. Genomic and expression profiling of chromosome 17 in 2008;27:910–20. https://doi.org/10.1038/emboj.2008.22. breast cancer reveals complex patterns of alterations and novel 36. Ahn R, Sabourin V, Bolt AM, Hebert S, Totten S, De Jay N, can- didate genes. Cancer Res. 2004;64:6453–60. et al. The Shc1 adaptor simultaneously balances Stat1 and Stat3 19. Rennhack JP, To B, Swiatnicki M, Dulak C, Ogrodizinski MP, activity to promote breast cancer immune suppression. Nat Zhang Y, et al. Integrated analyses of murine breast cancer Commun. 2017;8:14638. https://doi.org/10.1038/ncomms14638. models reveal critical parallels with human disease. Nat Com- 37. Wagner K, Wall RJ, St-Onge L, Gruss P, Wynshaw-Boris A, mun. 2019;10:3261. Garrett L, et al. Cre-mediated gene deletion in the mammary 20. Swiatnicki MR, Rennhack JP, Andrecheck ER. PTPRH Muta- gland. Nucleic Acids Res. 1997;25:4323–30. tions in NSCLC regulates EGFR phosphorylation. J Thorac 38. White DE, Kurpios NA, Zuo D, Hassell JA, Blaess S, Mueller U, Oncol. 2020;15:25. et al. Targeted disruption of beta1-integrin in a transgenic mouse 21. Xiao B, Zuo D, Hirukawa A, Cardiff RD, Lamb R, Sonenberg N, model of human breast cancer reveals an essential role in et al. Rheb1-Independent Activation of mTORC1 in Mammary mammary tumor induction. Cancer Cell. 2004;6:159–70. https:// Tumors Occurs through Activating Mutations in mTOR. Cell Rep. doi.org/10.1016/j.ccr.2004.06.025. 2020;31:107571. https://doi.org/10.1016/j.celrep.2020.107571. 39. Rao T, Ranger JJ, Smith HW, Lam SH, Chodosh L, Muller WJ. 22. Cai Y, Nogales-Cadenas R, Zhang Q, Lin J, Zhang W, O’Brien Inducible and coupled expression of the polyomavirus middle T K, et al. Transcriptomic dynamics of breast cancer progression in antigen and Cre recombinase in transgenic mice: an in vivo the MMTV-PyMT mouse model. 2017;18:1–14. model for synthetic viability in mammary tumour progression. 23. Dunant NM, Senften M, Ballmer-Hofer K. Polyomavirus Breast Cancer Res. 2014;16:R11. https://doi.org/10.1186/ middle-T antigen associates with the kinase domain of Src- bcr3603. related tyrosine kinases. J Virol. 1996;70:1323–30. 40. Gunther EJ, Belka GK, Wertheim GB, Wang J, Hartman JL, 24. Campbell KS, Ogris E, Burke B, Su W, Auger KR, Druker BJ, Boxer RB, et al. A novel doxycycline-inducible system for the et al. Polyoma middle tumor antigen interacts with SHC protein transgenic analysis of mammary gland biology. FASEB J. via the NPTY (Asn-Pro-Thr-Tyr) motif in middle tumor antigen. 2002;16:283–92. https://doi.org/10.1096/fj.01-0551com. Proc Natl Acad Sci USA. 1994;91:6344–8. https://doi.org/10. 41. Jones LM, Broz ML, Ranger JJ, Ozcelik J, Ahn R, Zuo D, et al. 1073/pnas.91.14.6344. STAT3 Establishes an immunosuppressive microenvironment 25. Whitman M, Kaplan DR, Schaffhausen B, Cantley L, Roberts TM. during the early stages of breast to promote tumor Association of phosphatidylinositol kinase with polyoma middle-T growth and metastasis. Cancer Res. 2016;76:1416–28. https:// competent for transformation. Nature. 1985;315:239–42. doi.org/10.1158/0008-5472.CAN-15-2770. 26. Su W, Liu W, Schaffhausen BS, Roberts TM. Association of 42. Hirukawa A, Smith HW, Zuo D, Dufour CR, Savage P, Bertos polyomavirus middle T with phospholipase C-gamma1. J Biol N, et al. Targeting EZH2 reactivates a breast cancer subtype- Chem. 1995;270:12331–4. specific anti-metastatic transcriptional program. Nat Commun. 27. Pallas DC, Shahrik LK, Martin BL, Jaspers S, Miller TB, 2018;9:2547. https://doi.org/10.1038/s41467-018-04864-8. Brautigan DL, et al. Polyoma small and middle T antigens and 43. Gross ET, Han S, Vemu P, Peinado CD, Marsala M, Ellies LG, SV40 small t antigen form stable complexes with protein phos- et al. Immunosurveillance and immunoediting in MMTV-PyMT- phatase 2A. Cell. 1990;60:167–76. induced mammary oncogenesis. Oncoimmunology. 2017;6: 28. Hunter T, Hutchinson MA, Eckhart W. Polyoma middle-T e1268310. https://doi.org/10.1080/2162402X.2016.1268310. antigen can be phosphorylated on tyrosine at multiple sites 44. Qin J, Yan L, Zhang J, Zhang WD. STAT3 as a potential ther- in vitro. EMBO J. 1984;3:73–80. apeutic target in triple negative breast cancer: a systematic 29. Ong SH, Dilworth S, Hauck-Schmalenberger I, Pawson T, Kiefer F. review. J Exp Clin Cancer Res. 2019;38. https://doi.org/10.1186/ ShcA and Grb2 mediate polyoma middle T antigen-induced endo- s13046-019-1206-z thelial transformation and Gab1 tyrosine phosphorylation. EMBO J. 45. Halaoui R, Rejon C, Chatterjee SJ, Szymborski J, Meterissian S, 2001;20:6327–36. https://doi.org/10.1093/emboj/20.22.6327. Muller WJ, et al. Progressive polarity loss and luminal collapse 30. Kaplan DR, Whitman M, Schaffhausen B, Raptis L, Garcea RL, disrupt tissue organization in carcinoma. Genes Dev. Pallas D, et al. Phosphatidylinositol metabolism and polyoma- 2017;31:1573–87. https://doi.org/10.1101/gad.300566.117. mediated transformation. PNAS. 1986;83:3624–8. 46. Allred DC, Mohsin SK, Fuqua SA. Histological and biological 31. Guy CT, Muthuswamy SK, Cardiff RD, Soriano P, Muller WJ. evolution of human premalignant breast disease. Endocr Relat Activation of the c-Src tyrosine kinase is required for the Cancer. 2001;8:47–61. https://doi.org/10.1677/erc.0.0080047. induction of mammary tumors in transgenic mice. Genes Dev. 47. Simond AM, Ling C, Moore MJ, Condotta SA, Richer MJ, 1994;8:23–32. https://doi.org/10.1101/gad.8.1.23. Muller WJ. Point-activated ESR1(Y541S) has a dramatic effect 32. Kim H, Laing M, Muller W. c-Src-null mice exhibit defects in on the development of sexually dimorphic organs. Genes Dev. normal mammary gland development and ERalpha signaling. 2020;34:1304–9. https://doi.org/10.1101/gad.339424.120. Oncogene. 2005;24:5629–36. https://doi.org/10.1038/sj.onc. 48. Kleiner D, Stetler-Stevenson W. Matrix metalloproteinases and 1208718. metastasis. Cancer Chemother Pharm. 1999;43:S42–S51. 33. Marcotte R, Smith HW, Sanguin-Gendreau V, McDonough RV, 49. Winer A, Adams S, Mignatti P. Matrix Inhi- Muller WJ. Mammary epithelial-specific disruption of c-Src bitors in Cancer Therapy: Turning Past Failures Into Future impairs cell cycle progression and tumorigenesis. Proc Natl Acad Successes. Mol Cancer Therapeutics. 2018;17:1147–55. https:// Sci USA. 2012;109:2808–13. https://doi.org/10.1073/pnas. doi.org/10.1158/1535-7163.MCT-17-0646. 1018861108. 50. Decock J, Hendrickx W, Thirkettle S, Gutiérrez-Fernández A, 34. Webster MA, Hutchinson JN, Rauh MJ, Muthuswamy SK, Anton Robinson SD, Edwards DR. Pleiotropic functions of the tumor- M, Tortorice CG, et al. Requirement for both Sch and phosphati- and metastasis-suppressing matrix metalloproteinase-8 in mam- dylinositol 39 kinase pathways in polyomavirus middle T-mediated mary cancer in MMTV-PyMT transgenic mice. Breast Cancer mammary tumorigenesis. Mol Cell Biol. 1998;18:2344–59. Res. 2015;17:38 https://doi.org/10.1186/s13058-015-0545-8. 490 S. Attalla et al.

51. Höckel M, Vaupel P. Tumor hypoxia: definitions and current 68. Lahlou H, Sanguin-Gendreau V, Zuo D, Cardiff RD, McLean clinical, biologic, and molecular aspects. J Natl Cancer Inst. GW, Frame MC, et al. Mammary epithelial-specific disruption of 2001;ume 93:266–76. https://doi.org/10.1093/jnci/93.4.266. the focal adhesion kinase blocks mammary tumor progression. 52. Majumdar AJ, Wong WJ, Simon MC. Hypoxia-inducible factors Proc Natl Acad Sci USA. 2007;104:20302–7. https://doi.org/10. and the response to hypoxic stress. Mol Cell. 2010;40:294–309. 1073/pnas.0710091104. 53. Jin M, Lee H, Park IA, Chung YR, Im SA, Lee KH, et al. 69. Gronbaek K, Hother C, Jones PA. Epigenetic changes in cancer. Overexpression of HIF1α and CAXI predicts poor outcome in APMIS. 2007;115:1039–59. early-stage triple negative breast cancer. Virchows Arch. 70. Esteller M. CpG island hypermethylation and tumor suppressor 2016;469:183–90. genes: a booming present, a brighter future. Oncogene. 54. Schwab LP, Peacock DL, Majumdar D, Ingels JF, Jensen LC, 2002;21:5427–40. https://doi.org/10.1038/sj.onc.1205600. Smith KD, et al. Hypoxia-inducible factor 1α promotes primary 71. Di Croce L, Helin K. Transcriptional regulation by Polycomb tumor growth and tumor-initiating cell activity in breast cancer. group . Nat Struct Mol Biol. 2013;20:1147–55. https:// Breast Cancer Res. 2012;14:R6. https://doi.org/10.1186/bcr3087. doi.org/10.1038/nsmb.2669. 55. Liao D, Corle C, Seagroves TN, Johnson RS. Hypoxia-inducible 72. Kleer CG, Cao Q, Varambally S, Shen R, Ota I, Tomlins SA, factor -1α is a key regulator of metastasis in a transgenic model of et al. EZH2 is a marker of aggressive breast cancer and promotes cancer initiation and progression. Cancer Res. 2007;67:563–72. neoplastic transformation of breast epithelial cells. Proc Natl 56. Creighton CJ, Li X, Landis M, Dixon JM, Neumeister VM, Acad Sci USA. 2003;100:11606–11. https://doi.org/10.1073/pna Sjolund A, et al. Residual breast cancers after conventional s.1933744100. therapy display mesenchymal as well as tumor-initiating fea- 73. Li X, Gonzalez ME, Toy K, Filzen T, Merajver SD, Kleer CG. tures. Proc Natl Acad Sci USA. 2009;106:13820–5. https://doi. Targeted overexpression of EZH2 in mammary gland disrupts org/10.1073/pnas.0905718106. ductal morphoenesis and causes epithelial hyperplasia. Am J 57. Ma J, Lanza DG, Guest I, Uk-Lim C, Glinskii A, Glinsky G, Pathol. 2009;175:1246–54. et al. Characterization of mammary cancer stem cells in the 74. Cai Y, Nogales-Cadenas R, Zhang Q, Lin JR, Zhang W, O’Brien MMTV-PyMT mouse model. Tumor Biol. 2012; https://doi.org/ K, et al. Transcriptomic dynamics of breast cancer progression in 10.1007/s13277-012-0458-4 the MMTV-PyMT mouse model. BMC Genomics. 2017;18:185 58. Asselin-Labat ML, Sutherland KD, Vaillant F, Gyorki DE, Wu https://doi.org/10.1186/s12864-017-3563-3. D, Holroyd S, et al. Gata-3 negatively regulates the tumor- 75. Cai Y, Lin JR, Zhang Q, O’Brien K, Montagna C, Zhang ZD. initiating capacity of mammary luminal progenitor cells and Epigenetic alterations to Polycomb targets precede malignant targets the putative tumor suppressor caspase-14. Mol Cell Biol. transition in a mouse model of breast cancer. Sci Rep. 2011;31:4609–22. https://doi.org/10.1128/MCB.05766-11. 2018;8:5535 https://doi.org/10.1038/s41598-018-24005-x. 59. Ni T, Li X, Lu N, An T, Liu Z, Fu R, et al. Snail1-dependent 76. Hirukawa A, Singh S, Wang J, Rennhack JP, Swiatnicki M, repression regulates expansion and activity of tumour-initiating Sanguin-Gendreau V, et al. Reduction of global H3K27me(3) cells in breast cancer. Nat Cell Biol. 2016;18:1221–32. https:// enhances HER2/ErbB2 targeted therapy. Cell Rep. 2019;29:249–57 doi.org/10.1038/ncb3425. e248. https://doi.org/10.1016/j.celrep.2019.08.105. 60. Dadi S, Chhangawala S, Whitlock BM, Franklin RA, Luo CT, 77. Davie SA, Maglione JE, Manner CK, Young D, Cardiff RD, Oh SA, et al. Cancer Immunosurveillance by Tissue-Resident MacLeod CL, et al. Effects of FVB/NJ and C57Bl/6J strain Innate Lymphoid Cells and Innate-like T Cells. Cell. backgrounds on mammary tumor phenotype in inducible nitric 2016;164:365–77. https://doi.org/10.1016/j.cell.2016.01.002. oxide synthase deficient mice. Transgenic Res. 2007;16:193–201. 61. Boyle ST, Faulkner JW, McColl SR, Kochetkova M. The che- https://doi.org/10.1007/s11248-006-9056-9. mokine receptor CCR6 facilitates the onset of mammary neo- 78. Lifsted T, Le Voyer T, Williams M, Muller W, Klein-Szanto A, plasia in the MMTV-PyMT mouse model via recruitment of Buetow KH, et al. Identification of inbred mouse strains harboring tumor-promoting macrophages. Mol Cancer. 2015;14:115. genetic modifiers of mammary tumor age of onset and metastatic https://doi.org/10.1186/s12943-015-0394-1. progression. Int J Cancer. 1998;77:640–4. https://doi.org/10.1002/ 62. Harbeck N, Penault-Llorca F, Cortes J, Gnant M, Houssami N, (sici)1097-0215(19980812)77:4<640::Aid-ijc26>3.0.Co;2-8. Poortmans P, et al. Breast cancer. Nat Rev Dis Prim. 2019;5:66. 79. Lukes L, Crawford NP, Walker R, Hunter KW. The origins of https://doi.org/10.1038/s41572-019-0111-2. breast cancer prognostic gene expression profiles. Cancer Res. 63. Butt SA, Sogaard LV, Ardenkjaer-Larsen JH, Lauritzen MH, 2009;69:310–8. https://doi.org/10.1158/0008-5472.CAN-08-3520. Engelholm LH, Paulson OB, et al. Monitoring mammary tumor 80. Hunter KW, Broman KW, Le Voyer T, Lukes L, Cozma D, progression and effect of tamoxifen treatment in MMTV-PyMT Debies MT, et al. Predisposition to efficient mammary tumor using MRI and magnetic resonance spectroscopy with hyper- metastatic progression is linked to the breast cancer metastasis polarized [1-13C] Pyruvate. Magn Reson Med. 2015;73:51–58. suppressor gene Brms1. Cancer Res. 2001;61:8866–72. 64. Qin L, Liao L, Redmond A, Young L, Yuan Y, Chen H, et al. 81. Park YG, Zhao X, Lesueur F, Lowy DR, Lancaster M, Pharoah The AIB1 oncogene promotes breast cancer metastasis by acti- P, et al. Sipa1 is a candidate for underlying the metastasis effi- vation of PEA3-mediated matrix metalloproteinase 2 (MMP2) ciency modifier locus Mtes1. Nat Genet. 2005;37:1055–62. and MMP9 expression. Mol Cell Biol. 2008;28:5937–50. https:// https://doi.org/10.1038/ng1635. doi.org/10.1128/MCB.00579-08. 82. Boyd NF, Lockwood GA, Byng JW, Tritchler DL, Yaffe MJ. 65. Dabrosin C, Palmer K, Muller WJ, Gauldie J. Estradiol promotes Mammographic densities and breast cancer risk. Cancer Epide- growth and angiogenesis in polyoma middle T transgenic mouse miol Biomark Prev. 1998;7:1133–44. mammary tumor explants. Breast Cancer Res Treat. 2003;78:1–6. 83. Martin LJ, Melnichouk O, Guo H, Chiarelli AM, Hislop TG, Yaffe 66. Cook RS, Garrett JT, Sanchez V, Stanford JC, Young C, MJ, et al. Family history, mammographic density, and risk of Chakrabarty A, et al. ErbB3 ablation impairs PI3K/Akt-depen- breast cancer. Cancer Epidemiol Biomark Prev. 2010;19:456–63. dent mammary tumorigenesis. Cancer Res. 2011;71:3941–51. 84. Gill JK, Maskarinec G, Pagano I, Kolonel LN. The association of 67. Jang I, Beningo KA. Integrins, CAFs and mechanical forces in mammographic density with ductal carcinoma in situ of the the progression of cancer. Cancers. 2019;11:721. https://doi.org/ breast: the Multiethnic Cohort. Breast Cancer Res. 2006;8:R30. 10.3390/cancers11050721. https://doi.org/10.1186/bcr150. Insights from transgenic mouse models of PyMT-induced breast cancer: recapitulating human breast cancer. . . 491

85. Provenzano PP, Inman DR, Eliceiri KW, Knittel JG, Yan L, factor in myeloid cells accelerates tumorigenesis. Nature. Rueden CT, et al. Collagen density promotes mammary tumor 2008;456:814–8. https://doi.org/10.1038/nature07445. initiation and progression. BMC Med. 2008;6:11. https://doi.org/ 100. Allen E, Jabouille A, Rivera LB, Lodewijckx I, Missiaen R, Steri 10.1186/1741-7015-6-11. V, et al. Combined antiangiogenic and anti-PD-L1 therapy sti- 86. Esbona K, Inman D, Saha S, Jeffert J, Schedin P, Wilke L, et al. mulates tumor immunity through HEV formation. Sci Transl COX-2 modulates mammary tumor progression in response to Med. 2017;9: https://doi.org/10.1126/scitranslmed.aak9679 collagen density. Breast Cancer Res. 2016;18:35. https://doi.org/ 101. Mazzieri R, Pucci F, Moi D, Zonari E, Ranghetti A, Berti A, 10.1186/s13058-016-0695-3. et al. Targeting the ANG2/TIE2 axis inhibits tumor growth and 87. Ostuni R, Kratochvill F, Murray PJ, Natoli G. Macrophages and metastasis by impairing angiogenesis and disabling rebounds of cancer: from mechanisms to therapeutic implications. Trends proangiogenic myeloid cells. Cancer Cell. 2011;19:512–26. Immunol. 2015;36:229–39. https://doi.org/10.1016/j.it.2015.02.004. https://doi.org/10.1016/j.ccr.2011.02.00. 88. Franklin RA, Liao W, Sarkar A, Kim MV, Bivona MR, Liu K, 102. Welford AF, Biziato D, Coffelt SB, Nucera S, Fisher M, Pucci F, et al. The cellular and molecular origin of tumor-associated et al. TIE2-expressing macrophages limit the therapeutic efficacy of macrophages. Science. 2014;344:921–5. https://doi.org/10.1126/ the vascular-disrupting agent combretastatin A4 phosphate in mice. J science.1252510. Clin Invest. 2011;121:1969–73. https://doi.org/10.1172/JCI44562. 89. Lin Y, Xu J, Lan H. Tumor-associated macrophages in tumor 103. Kesavan R, Frömel T, Zukunft S, Laban H, Geyer A, Naeem Z, metastasis: biological roles and clinical therapeutic applications. et al. Cyp2c44 regulates prostaglandin synthesis, lymphangio- J Hematol Oncol. 2019;12:76. https://doi.org/10.1186/s13045- genesis, and metastasis in a mouse model of breast cancer. Proc 019-0760-. Natl Acad Sci USA. 2020;117:5923–30. https://doi.org/10.1073/ 90. Ojalvo LS, King W, Cox D, Pollard JW. High-density gene pnas.1921381117. expression analysis of tumor-associated macrophages from 104. Wagner K-U. Models of breast cancer: quo vadis, animal mod- mouse mammary tumors. Am J Pathol. 2009;174:1048–64. eling? Breast Cancer Res. 2003;6. https://doi.org/10.1186/bcr723 https://doi.org/10.2353/ajpath.2009.080676. 105. Forrester E, Chytill A, Bierie B, Aakre M, Gorska AE, Sharid- 91. Lin EY, Li JF, Bricard G, Wang W, Deng Y, Sellers R, et al. Afshar A, et al. Effect of conditional knockout of the type II Vascular endothelial growth factor restores delayed tumor pro- TGFB receptor gene in mammary epithelia on mammary gland gression in tumors depleted of macrophages. Mol Oncol. development and polyomavirus middle T antigen induced tumor 2007;1:288–302. https://doi.org/10.1016/j.molonc.2007.10.003. formation and metastasis. Cancer Res. 2005;65:2296–302. 92. Weichand B, Popp R, Dziumbla S, Mora J, Strack E, Elwakeel 106. Arun G, Diermeier S, Akerman M, Chang KC, Wilkinson JE, E, et al. S1PR1 on tumor-associated macrophages promotes Hearn S, et al. Differentiation of mammary tumors and reduction in lymphangiogenesis and metastasis via NLRP3/IL-1beta. J Exp metastasis upon Malat1 lncRNA loss. Genes Dev. 2015;30:34–51. Med. 2017;214:2695–713. https://doi.org/10.1084/jem.2016039. 107. Vasiljeva O, Papazoglou A, Kruger A, Brodoefel H, Korovin M, 93. Welm AL, Sneddon JB, Taylor C, Nuyten DS, van de Vijver MJ, Deussing J, et al. Tumor cell-derived and macrophage derived Hasegawa BH, et al. The macrophage-stimulating protein path- cathepsin B promotes progression and lung metastasis of mam- way promotes metastasis in a mouse model for breast cancer and mary cancer. Cancer Res. 2006;66:5241–50. predicts poor prognosis in humans. Proc Natl Acad Sci USA. 108. Harris IS, Trekoar AE, Inoue S, Sasaki M, Gorrini C, Lee KC, 2007;104:7570–5. https://doi.org/10.1073/pnas.0702095104. et al. Glutathione and thioredoxin antioxidant pathways syner- 94. Maggiora P, Marchio S, Stella MC, Giai M, Belfiore A, De gize to drive cancer initiation and progression. Cancer Cell. Bortoli M, et al. Overexpression of the RON gene in human 2015;27:211–22. breast carcinoma. Oncogene. 1998;16:2927–33. https://doi.org/ 109. Li J, Karaplis AC, Huang DC, Siegel PM, Camirand A, Yang 10.1038/sj.onc.1201812. XF, et al. PTHrP drives breast tumor initiation, progression and 95. Eyob H, Ekiz HA, Derose YS, Waltz SE, Williams MA, Welm metastasis in mice and is a potential therapy target. JCI. AL. Inhibition of ron kinase blocks conversion of micro- 2011;121:4655–69. metastases to overt metastases by boosting antitumor immunity. 110. Pützer BM, Hitt M, Muller WJ, Emtage P, Gauldie J, Graham Cancer Discov. 2013;3:751–60. https://doi.org/10.1158/2159- FL. Interleukin 12 and B7-1 costimulatory molecule expressed 8290.CD-12-0480. by an adenovirus vector act synergistically to facilitate tumor 96. Ekiz HA, Lai SA, Gundlapalli H, Haroun F, Williams MA, regression. Proc Natl Acad Sci USA. 1997;94:10889–94. https:// Welm AL. Inhibition of RON kinase potentiates anti-CTLA-4 doi.org/10.1073/pnas.94.20.10889. immunotherapy to shrink breast tumors and prevent metastatic 111. Movahedi K, Schoonooghe S, Laoui D, Houbracken I, Waelput outgrowth. Oncoimmunology. 2018;7:e1480286. https://doi.org/ W, Breckpot K, et al. Nanobody-based targeting of the macro- 10.1080/2162402X.2018.1480286. phage mannose receptor for effective in vivo imaging of tumor- 97. Wculek SK, Malanchi I. Neutrophils support lung colonization assoiated macrophages. Cancer Res. 2012;72:4165–77. of metastasis-initiating breast cancer cells. Nature. 112. Shiao SL, Ruffell B, DeNardo DG, Faddegon BA, Park CC, 2015;528:413–7. https://doi.org/10.1038/nature16140. Coussens LM. TH2-Polarized CD4(+) T Cells and Macrophages 98. Ohs I, Ducimetiere L, Marinho J, Kulig P, Becher B, Tugues S. Limit Efficacy of Radiotherapy. Cancer Immunol Res. Restoration of natural killer cell antimetastatic activity by IL12 2015;3:518–25. https://doi.org/10.1158/2326-6066.CIR-14-0232. and checkpoint blockade. Cancer Res. 2017;77:7059–71. https:// 113. Nakasone ES, Askautrud HA, Kees T, Park JH, Plaks V, Ewald doi.org/10.1158/0008-5472.CAN-17-1032. AJ, et al. Imaging tumor-stroma interactions during chemotherapy 99. Stockmann C, Doedens A, Weidemann A, Zhang N, Takeda N, reveals contributions of the microenvironment to resistance. Cancer Greenberg JI, et al. Deletion of vascular endothelial growth Cell. 2012;21:488–503. https://doi.org/10.1016/j.ccr.2012.02.017.